Confusion Matrix setup in scikit-learn

[ouverz]

Hi!
I wanted to start off by saying that your tutorials and videos are really great! so clear and simple!

I've been working on a binary classification problem for my school with scikit-learn and I have been scratching my head in regards to how it displays the confusion matrix. For instance I have as output

 [ [30  5]
    [2 42] ]

I noticed by looking at the classification report that scikit learn by default outputs the negative class first. This leads me to understand that the first list is the negative class and that the second is the positive class. However, what I don't understand how to interpret what each number stands for as in TP, FP, TN, FN.

TN(30) FN (5)
FP(2) TP (42)

Is this a current representation of the input above?

Thanks a bunch!

[justmarkham]

Hi @ouverz, thanks for your kind comments!

Regarding your interpretation of the confusion matrix, you have the FP and FN backwards. It should be:

TN(30) FP(5)
FN(2) TP(42)

Scroll down to the Confusion Matrix section of this notebook for more information and another example: https://github.com/justmarkham/scikit-learn-videos/blob/master/09_classification_metrics.ipynb

Hope that helps!